How to work with Set Differences in Java using objects?

Question

I need to obtain a set difference between two sets/lists but the objects in these sets are constructed differently. I would like the set difference be calculated on only one field of these objects, how can I do this?

POJO o1 = new POJO();
o1.setName("Bean");
o1.setLocation("Boston");

POJO o2 = new POJO();
o2.setName("Bagel");
o2.setLocation("NYC");

POJO o3 = new POJO();
o3.setName("Bean");

List<POJO> p1 = new ArrayList<>();
p1.add(o1);
p1.add(o2);

List<POJO> p2 = new ArrayList<>();
p2.add(o3);

Collection<POJO> subList = CollectionUtils.subtract(p1, p2);
System.out.println("subList: "+subList.toString());

The output I see is : subList: [POJO(name=Bean)] i.e. o3. But for my use case, o1 and o3 are the same - how can I achieve this? I overrode the default equals() method as well but i don't think it is invoked.

I had to search to see where `CollectonUtils` was from. It appears to be from `Apache`. An overloaded version takes a predicate as the third argument for determining which values to subtract. Please update your tags to include the appropriate third party API reference. — WJS, Jan 23 '20 at 20:29
https://stackoverflow.com/questions/49324437/is-there-a-simple-way-in-java-to-get-the-difference-between-two-collections-usin/49332217#49332217 — Donald Raab, Jan 24 '20 at 02:18

score 4 · Accepted Answer · answered Jan 23 '20 at 20:59

You can achieve this with the help of overloaded subtract with predicate.

Override hashcode and equals in POJO class, (and toString for clean output)

@Override
public boolean equals(Object o) {
    if (this == o) return true;
    if (o == null || getClass() != o.getClass()) return false;
    POJO pojo = (POJO) o;
    return Objects.equals(name, pojo.name);
}

@Override
public int hashCode() {
    return name.hashCode();
}

Use Predicate with set as

Collection<POJO> subList = CollectionUtils.subtract(p1, p2, new Predicate<POJO>() {
    @Override
    public boolean evaluate(POJO object) {
        return p1.contains(object); // might need to make p1 final
    }
});

Result:

subList: [ Bagel ]

Note: Apache common does not support anything like Java 8 BiPredicate

This is incredible! I didn't know we could override the Predicate in the `subtract` method. Thanks a ton for this! — Saturnian, Jan 23 '20 at 21:43

score 4 · Answer 2 · answered Jan 24 '20 at 02:36

You can use a UnifiedSetWithHashingStrategy from Eclipse Collections:

POJO o1 = new POJO();
o1.setName("Bean");
o1.setLocation("Boston");

POJO o2 = new POJO();
o2.setName("Bagel");
o2.setLocation("NYC");

POJO o3 = new POJO();
o3.setName("Bean");

HashingStrategy<POJO> strategy = HashingStrategies.fromFunction(POJO::getName);

MutableSet<POJO> set1 =
        HashingStrategySets.mutable.with(strategy, o1, o2);
MutableSet<POJO> set2 =
        HashingStrategySets.mutable.with(strategy, o3);
MutableSet<POJO> difference = set1.difference(set2);

Assert.assertEquals("[POJO(name=Bagel)]", difference.toString());

This answer explains how UnifiedSetWithHashingStrategy works in more detail.

Note: I am a committer for Eclipse Collections.

How to work with Set Differences in Java using objects?

2 Answers2